Effect of distal perfusion on spinal cord blood flow during aortic cross-clamping and cerebrospinal fluid pressure elevation

Background: Permanent paraplegia is a rare but feared complication of both open and endovascular thoracoabdominal aortic repair. The rate of postoperative paraplegia varies depending on the extent of open repair, from 0.9% to 4.7%, even in expert centers (Coselli et al. 2016; Etz et al. 2015). Among the currently available protective adjuncts, distal perfusion (DP) and cerebrospinal fluid (CSF) drainage are one of the most widely used ones (Etz et al. 2015). The scientific evidence of DP is based on observational clinical studies with heterogenous patients and perioperative strategies, and few experimental works with various combinations of preventive techniques analyzed simultaneously (Rose et al. 1997; Winnerkvist et al. 2002). Aim of the study: The aim of the study was to evaluate the isolated effect of DP on regional spinal cord perfusion during aortic cross-clamping, and additional deliberate CSF pressure elevation in a large animal model. Additionally, we aimed to assess DP impact on paraspinous muscle perfusion, and evaluate the efficacy of collateral network near-infrared spectroscopy (cnNIRS) as a monitoring technique during DP. Methods: The study was performed in an acute large animal model (8 juvenile female German landrace pigs) via upper left lateral thoracotomy in the 3rd intercostal space, and retroperitoneal access. Distal perfusion was performed using partial cardiopulmonary bypass (CPB) with target perfusion pressure of 60 mmHg. Arterial lines of CPB were placed into the descending thoracic and abdominal aorta, and the venous line – into the pulmonary artery. Lumbar puncture at the L3-L4 level was performed in order to perform plasma injection during CSF pressure elevation stage. Spinal cord and paraspinous muscle regional perfusion was evaluated using microspheres injections (a total of 6 colors) at four experimental time-points: on running CPB (baseline), 5 minutes after proximal aortic cross-clamping, 5 minutes after abdominal aortic cross-clamping and initiation of DP, and after 15 minutes of manually increased (tripled) CSF pressure. During the DP, proximal and distal blood flows were evaluated separately with two microsphere colors injected simultaneously via CPB arterial lines. For the analysis, the spinal cord was divided into three segments: upper (C1-T7), mid- (T8-L2) and lower (L3-S). Paraspinous muscle perfusion and oxygenation were assessed at 4 levels: mid- and lower thoracic, upper and lower lumbar levels. At the end of each experiment, the whole spinal cord and 2 cm3 samples of paraspinous muscles corresponding to the cnNIRS levels, were harvested. Results: Spinal Cord Perfusion: In the upper spinal cord, statistically significant changes of regional perfusion were observed both after DTA cross-clamping (decrease to 62% from baseline), and after distal aortic cross-clamping with initiation of DP (increase to 156% in proximal and decrease to 5% from baseline in distal flow). These were followed by a significant drop of proximal spinal cord perfusion (from 152% back to 102%, p = 0.038), and some increase of distal perfusion values (from 5% to 19%, however not reaching statistical significance) during the CSF pressure elevation stage. In the mid-spinal cord, a notable decrease of perfusion was observed after proximal aortic cross-clamping (to 27%, p = 0.025). The initiation of DP was not associated with any notable changes in proximal and distal perfusion values. Afterwards, a decrease of proximal and distal perfusion values (from 33% to 13% in proximal, and from 24% to 10% in distal perfusion) was observed during CSF pressure elevation stage. These changes were, however, not statistically significant. Lower spinal cord measurements showed, similarly to mid-segment, a decrease in perfusion after DTA cross-clamping (to 14% from baseline, p = 0.001). Initiation of DP led to normalization of proximal perfusion of the lower spinal cord (to 96% from baseline). At the same time, it was associated with extreme hyperperfusion due to distal perfusion (up to 480% from baseline). The tripling of CSF pressure resulted in decrease of both proximal (from 96% to 59%, p = 0.131) and distal (from 480% to 468%, p = 0.999) perfusion rates. Paraspinous muscle perfusion: The analysis of paraspinous muscle (i.e. collateral network) perfusion values revealed few statistically significant changes. Proximal aortic cross-clamping resulted in a decrease of paraspinous muscle perfusion (not reaching statistical significance). The least perfused were lower thoracic and upper lumbar segments. Neither initiation of DP, nor CSF pressure elevation were associated with any statistically significant changes in paraspinous muscles perfusion at any of the analyzed levels, except the lower lumbar one. Here, the distal perfusion increased from 7% to 27% from baseline during DP, and from 27% to 60% during the CSF elevation stage (p = 0.034). cnNIRS: Continuous cnNIRS monitoring did not reveal any notable changes at the mid-thoracic level. At the other three levels, the values decreased after DTA cross-clamping (p < 0.001 according to ANOVA). At the lower thoracic level, the tissue oxygenation values crossed the 70% from baseline ischemic threshold after initiation of DP. CSF pressure elevation did not have any influence on cnNIRS values at any level. Discussion: Comparison of the present experiment with the previously published studies is limited due to discrepancy in experimental sequences, analyzed segments and possible effects of other protective adjuncts used in the studies. However, in the initial phase of the experiment, the decrease of blood flow in all the spinal cord segments, was similar to the previously published works (Brattli et al. 2007; von Aspern et al. 2020). These changes were used as a second, “ischemic” baseline during the present study. Initiation of the DP led to limited or no perfusion increase of spinal cord perfusion in upper (C1-T7) and mid-spinal cord (T8-L2). And, if in the upper spinal cord this could be compensated by increased proximal flow, the mid-spinal cord was the least protected segment. At the same time, it was associated with extreme hyperperfusion (due to distal flow) of the spinal cord in the lumbar segment (L3-S), which is a known risk factor of spinal cord injury itself (Bower et al. 1989; Gallagher et al. 2019). The CSF pressure elevation resulted in further spinal cord tissue perfusion decrease, as previously reported by Haunschild et al. in experiments without aortic cross-clamping and DP (Haunschild et al. 2020). Although these changes were statistically significant only at the upper spinal cord level, they resulted in a pronounced reduction of proximal perfusion also in the other two spinal cord segments. Similarly, the decrease was observed also in distal perfusion in the mid- and lower spinal cord. Summarizing these findings, one would suggest that not only did DP (with 60 mmHg pressure) not lead to adequate protection of the mid-spinal cord during aortic cross-clamping, but it also was not able to protect it in the presence of increased CSF pressure. One also needs to point out, that although elevated CSF pressure led to some decrease of distal flow in lumbar segment, it did not eliminate the hyperperfusion of the spinal cord. In paraspinous muscle perfusion, as previously reported by von Aspern and colleagues, the perfusion reduction was more prominent in the lower thoracic and upper lumbar segments, which corresponds with the spinal cord regional perfusion results. During the next stages, almost no changes were observed in paraspinous muscles perfusion. The exclusion was the lower lumbar level, where some increase of distal perfusion was observed, however not reflecting the hyperperfusion of the spinal cord at this level. As a reflection of collateral perfusion, collateral network oxygenation monitoring using cnNIRS demonstrated limited changes. The most pronounced decrease of oxygenation was observed after aortic cross-clamping, thus following the pattern reported by von Aspern (von Aspern et al. 2020). However, at lower thoracic level, the values did cross the 70% ischemic threshold after initiation of DP, signaling ischemia. Similarly to paraspinous muscle perfusion, cnNIRS was not able to reflect the hyperperfusion of the distal spinal cord. Conclusions: The present study points out, that DP during open thoracoabdominal aortic repair should be managed with caution. It was shown that DP with stable unadjusted perfusion pressure of 60 mmHg does not provide adequate protection at the mid-thoracic level of the spinal cord and could not counteract CSF pressure elevation. At the same time, it may be associated with hyperperfusion of its distal segment. Distal perfusion, both with normal and elevated CSF pressure, did not lead to any significant changes in paraspinous muscles perfusion, except the lower lumbar segment. However, the lowest perfusion values were observed around the mid-spinal cord (the most vulnerable) area. Moreover, despite the fact that cnNIRS was able to reflect severe spinal cord ischemia, it did not reveal the spinal cord hyperperfusion. Further studies, including chronic animal experiments, are required for precise evaluation of DP in various pressure modes, and its ability to counteract the elevated CSF pressure.:Table of contents III List of abbreviations V 1 Introduction 1 1.1 Anatomy of the aorta 1 1.2 Descending thoracic and thoracoabdominal aortic pathology 2 1.3 Open surgical and endovascular treatment of thoracic and thoracoabdominal aortic pathology 4 1.4 Postoperative spinal cord injury 7 1.5 Spinal cord anatomy 9 1.6 Spinal cord blood supply: collateral network concept 10 1.7 Perioperative and adjunctive strategies to prevent spinal cord injury 12 1.8 Swine as an experimental model for spinal cord injury research 15 2 Aim of the study 16 3 Materials 17 3.1 Experimental materials 17 3.1.1 Devices 17 3.1.2 Expendable materials and instruments 18 3.1.3 Medications and chemicals 20 3.2 Laboratory materials 22 3.2.1 Devices 22 3.2.2 Expendable materials and instruments 23 3.2.3 Chemicals 25 3.3 Software 26 4 Methods 27 4.1 Experimental model 27 4.1.1 Experimental animals 27 4.1.2 Anaesthesia 28 4.1.3 Surgical approach and experimental sequence 29 4.1.4 Tissue harvesting and preparation 33 4.2 Analysis during the experiment 33 4.2.1 Microsphere measurements 33 4.2.2 Collateral network near-infrared spectroscopy 38 4.2.3 Histopathological assessment 39 4.2.4 Statistical analysis 42 5 Results 43 5.1 Vital parameters during the experiment 43 5.2 Spinal cord regional perfusion 45 5.3 Collateral network regional perfusion 50 5.4 Collateral network oxygenation 54 5.5 Relationships between regional perfusion and oxygenation values 56 5.6 Histopathological assessment 60 6 Discussion 62 6.1 Discussion of vital parameters during the experiment 64 6.2 Discussion of spinal cord regional perfusion 66 6.3 Discussion of collateral network regional perfusion 71 6.4 Discussion of collateral network oxygenation 73 6.5 Discussion of the relationships between regional perfusion and 75 oxygenation values 6.6 Discussion of histopathological results 76 6.7 Conclusions 77 6.8 Limitations of the study 78 7 Summary 81 8 References 86 9 Figure legends 103 10 Table legends 105 Acknowledgements 106 Declaration about the independent work for dissertation 107 Curriculum vitae 10